J. Chem. Eng. Data 1988, 33,321-327 (5) Eng, R.; Sandler, S.I.J . Chem. Eng. Data 1984, 29, 156. (6) Hauschild, T.; Wu, H. S.;Sandler, S. I.J . Chem. Eng. Data 1987, 32, 226. (7) Boubllk, T.; Fried, V.; Hala, E. The Vapor Pressures of Pure Substances; Eisevier: Amsterdam, 1984. (8) Wilhok, R. C.; Zwolinski. B. J. Handbook of Vapor Pressures and Heats of Vaporization of Hydrocarbons and Related Compounds. API44-TRC Publications in Science and Engineerlng No. 101; Texas A&M University: College Station, TX, 1971. (9) Engineering Science Data Unit. Vapor Pressures and Critical Points of LiquMs X . C7 to C18 Aliphatic Amines; London, 1976. (10) TRC Thermodynamic Tables Non-hydrocarbons ; Thermodynamics Research Center; The Texas ABM University System: College Station, TX, 1965; k-5380.
-
321
(11) Eng, R. M.S. Thesis, University of Delaware, 1985. (12) Hayden, J. 0.; OConnell, J. P. Ind. Eng. Chem., Process Res. Dev. 1975, 14, 209. (13) Yen, L. C.; Woods, S. S.AIChEJ. 1966, 12, 95. (14) Casanova, C.; Rey, J.; Cobos, J. C. Int. DATA Ser., Sel. Data Mixtures, Ser. A 1986, 102. (15) Reid, R. C.; Prausnitz, J. M.; Poling, B. E. The Properties of gases & Llqukls, 4th ed.; McGraw-Hill: New York, 1987.
Received for review July 7, 1987. Accepted March 8, 1988. The research reported here was supported, in part, by the National Science Foundation Grant CBT-8612285 to the University of Delaware, and a grant from the Chevron Oil Field Research Co.
Phase Equilibria in Ternary Systems with Carbon Dioxide, Water, and Carboxylic Acids at Elevated Pressures Athanassios 2. Panaglotopoulos, * * + Richard C. Willson, and Robert C. Reid Department of Chemical Engineering, Massachusetts Institute of Technology, Cambridge, Massachusetts 02 739
Vapor-llquld and vapor-llquld-liquid equlllbrlum composltlons have been measured for the ternary systems CO, H,O acetlc acid at 313 and 333 K, CO, H,O proplonlc acld at 313 K, and CO, H,O n-butyrlc acld at 313 K, at pressures between 2.0 and 20.0 MPa. For all systems, a three-phase equlllbrlum region was observed at pressures comparable to the crltlcal pressure of pure carbon dioxide. The extent of the three-phase equlllbrlum reglon Increases wlth the hydrocarbon chaln length of the acld, whlle the pressure for the appearance of three phases also Increases wlth temperature. The distrjbutlon coefflclent of the aclds between the supercrltlcal fluld phase and the aqueous phase at pressures above the crltlcal pressure of pure carbon dioxlde Increases wlth the hydrocarbon chaln length of the acid: hlgher molecular welght acids were found to be preferentlally partittoned into the supercrltlcal fluid phase. The experlmental data for the ternary systems were correlated by using a cublc equatlon of state and a recently proposed density-dependent mlxlng rule wlth model parameters derived from blnary data only.
+
+
+
+
+
+
Introducllon Methods of using a fluid above its critical temperature and pressure as a solvent for the separation of mixtures of components have ,received considerable attention recently ( 7 , 2). The use of supercritical solvents for the energy-efficient recovery of low molecular weight organic compounds from aqueous solutions has been recently proposed by several investigators. Paulaitis et al. ( 3 ) ,McHugh et al. ( 4 ) ,and Gilbert and Paulaitis (5)investigated the recovery of ethanol with carbon dioxide, ethylene, and ethane as solvents. Kuk and Montagna (6) presented results for the separation of ethanol and 2-propanol from aqueous solutions using supercritical carbon dioxide. Radosz ( 7 ) and Paulaitis et al. (8, 9) have determined phase equilibria for the system 2-propanol water carbon dioxide. Fleck (70) presented results for the extraction of n -propyl alcohol using a variety of supercriticalfluids. Panagiotopoulos and Reid ( 7 7 , 72) have determined phase
+
+
'Current address: School of Chemical Engineerlng, Corneii University, Ithaca, NY 14853.
equilibria for ternary systems with supercritical carbon dioxide, water, and acetone or 1-butanol. There have been relatively few experimental investigations of equilibria between aqueous solutions of low molecular weight organic acids 'and supercritical fluids. Previous measurements include those of Snedeker (73)for the system carbon dioxide water acetic acid and Elgin and Weinstock (74) for the acetic acid and the corresponding system C2H4 -k water system with propionic acid. Francis ( 7 5 ) measured phase equilibria for the system subcritical carbon dioxide water acetic acid. Shimshick ( 1 6 ) has presented data for the extraction of acetic and n -butyric acid from aqueous solutions using supercritical carbon dioxide. The objective of the present work was the determination of the phase equilibrium behavior for the ternary systems supercritical carbon dioxide water one of the first three straight-chain aliphatic acids.
+
+
+
+
+
+
Experlmental Sectlon Equipment and Procedures. The experimental apparatus used was a high-pressure visual cell with external recirculation of all phases present. A detailed description of the equipment and operating procedures is given elsewhere ( 7 7 ) . Compositions were measured by gas chromatography with on-line sampling. The chromatographic response factors for the nonvolatile components were determined by analyzing standard mixtures prepared gravimetrically. The relative response factor for carbon dioxide was determined by measuring equilibrium data for mixtures for which data are available in the literature (78-20). The stability of the relative response factors was monitored by periodic injection of reference mixtures of the liquid components. An additional check for the validity of the chromatographic analysis was provided by the agreement of the results at high pressures with the known behavior of the binary CO, HO , system at low concentrations of organic acid. The validation of the experimental procedures and comparisons with literature data is given in ref 77. I t was determine8 that the composition of the less-dense supercritical fluid phase could not be reproducibly determined at lower pressures (56.0 MPa), primarily because of the small mass of sample injected. These data were not considered reliable and were not included in the data tables. Materials. Analytical grade acetic acid (supplied by Mallinckrodt), propionic acid (Sigma Chemical Co.), and n -butyric
+
0021-9568/88/1733-0321$01.50/00 1988 American Chemical Society
322 Journal of Chemical and Engineering Data, Vol. 33, No. 3, 1988 acid (Sigma) and glass-distilled water (Mallinckrodt) were used as received. Gas chromatographic analysis of the acids indicated a purity in the 99.5-99.7 wt % range, with water being the major contaminant. Since ternary data with water were to be measured, no effort was made to desiccate the acids. High purity CO, (99.99%, Matheson Gas Products) was used as received. Correlation For the correlation of the experimental results we used a cubic equation of state based on the Peng-Robinson form (27), with a new densitydependent mixing rule. A detailed discussion of the model used is given in ref 72. The mixing rule involves two interaction parameters per binary system, k = k , , and A,* = -A,,. The second parameter, A,, describes the deviation of the mixture from the quadratic mixing rule at high densities. The pure component parameters for the subcritical components (H,O and the acids) were determined by using a technique proposed by Panagiotopoulos and Kumar (22). The parameters thus obtained exactly reproduce vapor pressure and liquid density for the components of interest at each temperature. For the supercritical component (CO,), the conventional acentric factor correlation (27) was used with values for the critical constants taken from Reid et al. (23).
,
/
\r
=
c.
r
carbon dloxlde
woter
er
corboi dioxide
wote-
Results and Discussion The primary experimental results for the equilibrium compositions measured for the three systems are given in Tables I-IV. Each line in the tables gives the concentration of the coexisting phases (two or three) for the systems measured. Data from one to three chromatographic analyses (usually two) at each point were averaged to result in the reported concentration. The average standard deviation (reproducibility) for the set of analyses was f0.003 in mole fraction with a maximum deviation between duplicate analyses of f0.02. The absolute accuracy of the results is more difficult to ascertain. The uncertainty in the chromatographic response factors (77) is the main source of possible systematic errors. I f these uncertainties are taken into account, the accuracy of the reported mole fractions may be estimated to be f0.005 or f2% of the value reported, whichever is greater. The pressure inside the cell during the time required for the analyses (15-30min) was normally constant to within f 1.O% of the absolute value for the pressure; the pressure reported in the table is the average value. For the lower pressures, no data are available for the composition of the carbon dioxide-rich phase because of the difficulties in sampling mentioned in the previous section. The corresponding data are indicated by an “x” in the data tables. A full listing of the raw (not averaged) experimental results is available in Panagiotopoulos ( 77) and Willson (24);individual points are shown in Figures 1 and 2 and may help in estimating the reproducibility of the measurements. In Tables I-IV, the phases are listed in order of decreasing density. Generally, phase 1 is the water-rich phase, phase 3 is the C0,-rich phase, and phase 2 has an intermediate composition. However, this designation cannot be unique because of the transformation of phases into each other as pressure is changed. I n the tables, an attempt was made to be consistent and to indicate the evolution of the phase diagram with pressure. No exactly comparable results are available from the literature for any of the ternary systems studied in this work. A few data points from ref 13 are reported at similar conditions and agree with our results within the stated accuracy. An additional check on the accuracy of the results is provided by the reported concentrations of CO, in the aqueous phase and water in the CO, phase for low mole fraction of organic acid, extrapolated
WG
Figure 1. Phase equilibrium behavior for the system water-n-butyric acid-carbon dioxide at 313 K. (M) and (-) measured tie lines: (A) measured three-phase equilibrium compositions:(-) predicted tie lines. to zero concentration of acid, which agree with the accepted literature values from Wiebe and Gaddy (78) and Wiebe ( 7 9 ) . A typical evolution of the phase equilibrium behavior with pressure is given in Figure 1 for the system CO, HO , n-butyric acid at 313 K. At low pressures (P = 2.0 MPa) the diagram represents equilibrium between a low-density gas phase and a completely miscible liquid mixture. As pressure is increased, the dissolution of carbon dioxide in the liquid phase resutts in a split of the liquid into two phases, one richer in water and the other higher in acid concentration. The three-phase region is represented in Figure 1 as a triangular open area in the diagram for 4.0, 6.0, and 8.0 MPa. The physical picture that underlies this behavior, as first described by Elgin and Weinstock ( 7 4 ) ,is the “salting out” effect of a supercritical fluid on an aqueous solution of an organic compound. As pressure is increased, dissolution of the supercritical fluid in the aqueous phase results in a phase split at a lower critical solution pressure (which varies with temperature). As pressure is further increased, the second liquid phase and the supercritical phase increasingly similar and merge at an upper critical solution pressure, above which only two phases coexist at equilibrium. A three-phase equilibrium region was observed for all systems studied, as indicated in the data tables. The extent of the three-phase region (defined by the difference between the upper and lower critical solution pressure) becomes larger as the hydrocarbon chain length of the acid is increased. I t is of the order of 0.1 MPa for acetic acid at both 313 and 333 K, 0.5 MPa for propionic acid, and greater than 4 MPa for n-butyric acid. The three-phase equilibrium region moves to higher pressures as temperature is increased for the acetic acid system. The shape and extent of the two-phase coexistence region for all the systems studied were relatively insensitive to changes
+
+
Journal of Chemical and Engineering Data, Vol. 33, No. 3, 7988 323
Table I. Experimental Results for the System Water-Acetic Acid-Carbon Dioxide at 313 K mole fractions phase 2
phase 1 P,MPa 2.09 1.94 2.09 2.11 2.08 2.01 1.97 1.93 2.01 3.96 3.99 4.01 4.01 3.97 3.88 3.99 3.99 5.90 5.91 5.88 5.88 5.94 5.99 5.88 6.95 7.01 6.98 7.00 6.88 6.88 6.93 7.49 7.78 7.79 7.92 7.86 3-phase 7.80 3-phase 8.08 8.00 8.01 8.02 9.97 10.13 10.27 9.98 9.90 10.00 10.00 15.18 14.80 14.44 14.77 14.68 14.70 15.14 19.55 17.60 20.10 19.65 20.35 20.00 19.73
H20 0.011 0.026 0.078 0.158 0.359 0.565 0.778 0.891 0.944 0.008 0.068 0.139 0.318 0.537 0.765 0.877 0.933 0.004 0.043 0.098 0.276 0.511 0.877 0.927 0.003 0.024 0.062 0.230 0.494 0.702 0.748 0.001 0.024 0.030 0.571 0.749 0.868 0.925 0.960 0.472 0.575 0.683 0.785 0.874 0.931 0.962 0.402 0.550 0.693 0.772 0.879 0.926 0.963 0.408 0.540 0.709 0.783 0.881 0.927 0.958
acid 0.826 0.829 0.787 0.731 0.573 0.399 0.205 0.099 0.047 0.661 0.643 0.620 0.528 0.387 0.204 0.104 0.052 0.425 0.452 0.474 0.469 0.367 0.096 0.053 0.263 0.282 0.324 0.401 0.361 0.231 0.200 0.163 0.116 0.108 0.308 0.192 0.098 0.051 0.020 0.351 0.296 0.232 0.159 0.093 0.043 0.016 0.368 0.309 0.225 0.170 0.088 0.045 0.014 0.360 0.312 0.212 0.159 0.085 0.046 0.017
coz
H20
+
+
+
phase 3
COZ
0.163 0.145 0.135 0.111 0.069 0.037 0.017 0.010 0.009 0.331 0.289 0.241 0.154 0.076 0.032 0.019 0.015 0.571 0.506 0.427 0.255 0.122 0.027 0.020 0.734 0.694 0.613 0.369 0.145 0.066 0.052 0.836 0.860 0.862 0.121 0.060 0.034 0.024 0.020 0.177 0.130 0.085 0.056 0.033 0.027 0.023 0.231 0.140 0.082 0.058 0.033 0.029 0.023 0.232 0.148 0.079 0.058 0.034 0.027 0.024
H20 X X X X X X
X X X X
acid X X X
X X X X X X
co2 X X X X X
X X X X X
X
X X
X
X
X
X
X
X
X
X
X
X
X
X X X X X X X
X X
X X
X X X
X
X
X X X
X
X
X
X
X
X
X
X
X X
0.979
0.000
0.021
X
X
X
0.000
0.017
0.983
X X
X X
X X
0.001
0.017
0.983
X
X
X
0.015 0.009 0.012 0.008 0.015 0.014
0.985 0.990 0.987 0.992 0.984 0.985
0.001 0.007 0.032
0.080 0.103 0.132
0.919 0.891 0.835
0.000 0.001 0.002 0.000 0.001 0.001
X
X
X
X
X
X
0.004 0.001 0.000 0.063 0.022 0.019
0.006 0.001 0.003 0.347 0.162 0.157
0.990 0.998 0.998 0.590 0.817 0.825
X
X
X
X
X
X
0.012 0.003 0.067 0.026 0.012 0.025 0.005 0.005 0.002 0.069 0.034 0.014 0.015 0.006 0.004 0.003
0.006 0.001 0.266 0.186 0.103 0.074 0.030 0.008 0.001 0.270 0.193 0.108 0.071 0.029 0.012 0.001
0.982 0.997 0.667 0.788 0.885 0.901 0.965 0.987 0.997 0.661 0.772 0.878 0.913 0.965 0.984 0.997
in pressure and temperature above the critical pressure of the pure supercrklcal fluid, for the range of pressures between 10 and 20 MPa for the two temperatures studied. This is in agreement with the results of previous investigators for the ethanol water carbon dioxide system (5); it may not be correct when the system behavior at the high-pressure range is complicated by the presence of three-phase equilibrium regions. An example of this latter case is provided by the 2propanol water carbon dioxide system, which has been studied by Paulaitis et al. (8, 9) and Radosz (7).
+
acid
From the point of view of practical applications of the phase equilibrium behavior demonstrated here, the crucial parameters are the selectivity of the supercrkical fluid phase for the organic compound over water, and the ratio of acid mole fraction in the supercritical fluid phase to that in the aqueous phase, or distribution coefficient. A pronounced effect of the hydrocarbon chain length In a homologous series of organic compounds on the distribution coefficient and the selectivity of the supercritical solvent for the organic compound over water was observed. This can be best described by comparing the results for the
324 Journal of Chemical and Engineering Data, Vol. 33, No. 3, 1988
Table 11. Experimental Results for the System Water-Acetic AcidCarbon Dioxide at 333 K mole fractions phase 1
P, MPa 2.07 2.04 2.02 2.03 2.02 2.03 2.06 2.09 1.97 3.98 4.02 3.96 3.81 4.11 3.84 4.04 3.95 3.89 5.98 6.16 5.86 5.92 6.01 6.07 6.05 6.06 5.99 6.07 8.16 7.97 8.01 7.91 7.92 7.97 7.95 7.90 7.88 8.97 9.92 9.93 3-phase 10.24 9.98 9.91 10.17 10.20 10.30 10.25 10.20 13.21 10.91 10.88 12.08 11.02 12.07 12.18 11.04 12.04 11.06 12.04 11.95 14.83 15.04 14.94 15.05 14.90 14.79 14.79 15.05 18.40 19.96 20.00 19.63 19.85 19.30 19.75 19.64
H,O 0.012 0.139 0.415 0.467 0.564 0.680 0.860 0.949 0.978 0.009 0.124 0.389 0.457 0.554 0.662 0.865 0.940 0.972 0.009 0.106 0.356 0.433 0.435 0.529 0.649 0.859 0.935 0.968 0.007 0.084 0.331 0.413 0.502 0.636 0.850 0.930 0.964 0.007 0.212 0.295 0.389 0.488 0.630 0.843 0.930 0.963 0.341 0.366 0.419 0.472 0.510 0.544 0.573 0.634 0.644 0.857 0.929 0.960 0.315 0.530 0.580 0.673 0.859 0.932 0.934 0.962 0.381 0.456 0.525 0.595 0.661 0.858 0.933 0.962
acid 0.888 0.776 0.539 0.498 0.408 0.299 0.132 0.045 0.017 0.794 0.696 0.516 0.470 0.393 0.300 0.118 0.049 0.019 0.625 0.591 0.496 0.447 0.446 0.383 0.292 0.116 0.050 0.019 0.441 0.477 0.448 0.425 0.378 0.287 0.120 0.050 0.019 0.342 0.389 0.398 0.393 0.360 0.277 0.122 0.048 0.019 0.375 0.377 0.375 0.352 0.340 0.317 0.300 0.271 0.264 0.108 0.047 0.020 0.381 0.322 0.295 0.240 0.105 0.043 0.042 0.017 0.367 0.348 0.314 0.283 0.247 0.099 0.041 0.017
phase 2 COZ 0.099 0.085 0.046 0.035 0.028 0.021 0.008 0.006 0.005 0.197 0.180 0.096 0.073 0.052 0.038 0.018 0.011 0.010 0.366 0.303 0.149 0.120 0.120 0.089 0.059 0.025 0.016 0.014 0.552 0.440 0.221 0.162 0.120 0.077 0.030 0.019 0.017 0.652
0.399 0.307 0.218 0.152 0.093 0.035 0.022 0.018 0.284 0.257 0.206 0.176 0.150 0.140 0.127 0.095 0.092 0.035 0.024 0.020 0.304 0.148 0.125 0.087 0.036 0.025 0.025 0.021 0.252 0.196 0.160 0.122 0.092 0.043 0.025 0.022
HZO
acid
phase 3 COZ
H20 X X X X
X X X X
X X X X X X
X X
X X X X X
X
X X X X X X
X X X X X
X X
X X
X
X X
X X
X X X X
0.780 0.678 0.599
X X X
X X X
X X X X
X X
0.217 0.274 0.311
X
COZ X X X X X
X
X
0.004 0.048 0.090
acid X X X X X X
X
X
X
X X
X X
X X
X
X
X
0.001 0.009 0.103 0.087 0.100 0.003
0.064 0.083 0.049 0.030 0.027 0.026
0.936 0.908 0.849 0.883 0.874 0.971
X
X
X X
X X
0.006 0.001 0.001 0.007 0.009 0.001 0.001 0.001 0.004
0.000 0.056 0.070 0.067 0.073 0.026 0.024 0.019 0.016
0.994 0.944 0.929 0.926 0.918 0.973 0.974 0.980 0.980
X
X
X X
X X
0.004 0.065 0.047 0.022 0.028 0.020 0.031 0.023 0.034 0.034 0.052
0.004 0.242 0.176 0.100 0.135 0.060 0.094 0.094 0.057 0.073 0.010
0.993 0.693 0.778 0.878 0.837 0.920 0.875 0.882 0.909 0.894 0.938
X
X
X
0.004 0.069 0.027 0.022 0.012
0.000 0.247 0.139 0.117 0.079
0.996 0.683 0.834 0.861 0.909
X
X
X
0.010 0.006 0.006 0.073 0.050 0.042 0.024 0.017
0.003 0.003 0.001 0.248 0.205 0.167 0.132 0.102
0.988 0.991 0.993 0.679 0.745 0.791 0.843 0.881
X
X
X
0.007 0.006
0.007 0.001
0.986 0.994
Journal of Chemical and Engineering Data, Vol. 33, No. 3, 1988 325
Table 111. Experimental Results for the System Water-Propionic Acid-Carbon Dioxide at 313 K mole fractions ~
phase 1
P, M P a 1.96 2.00 2.04 1.99 2.00 2.03 2.16 1.99 2.05 4.00 3.89 3.98 4.04 4.10 4.17 4.15 4.06 4.00 5.96 6.06 5.99 6.04 6.02 6.03 6.04 6.03 5.92 7.99 8.02 7.79 3-phase 7.81 3-phase 7.98 8.22 8.26 10.12 10.44 10.33 10.17 10.30 15.01 15.36 14.97 15.01 15.49 14.70 15.19 14.73
H20 0033 0.196 0.361 0.636 0.741 0.817 0.903 0.950 0.981 0.016 0.136 0.339 0.600 0.715 0.804 0.897 0.944 0.974 0.015 0.102 0.257 0.538 0.690 0.779 0.883 0.938 0.964 0.006 0.009 0.667
acid 0.786 0.651 0.563 0.322 0.243 0.168 0.087 0.045 0.011 0.619 0.592 0.466 0.295 0.233 0.166 0.083 0.039 0.013 0.393 0.403 0.370 0.282 0.227 0.167 0.087 0.039 0.014 0.058 0.062 0.201
phase 3
phase 2
coz
H2O
acid
coz
HZO
0.181 0.153 0.076 0.042 0.016 0.015
X X X X
0.010
X
0.005 0.008 0.365 0.271 0.195 0.105 0.052 0.030 0.020 0.017 0.012 0.592 0.495 0.374 0.179 0.083 0.055 0.031 0.023 0.022 0.936 0.929 0.132
X X
X X
X X X X
X X X
X X
acid X X
X X
X X
X
X X
X X X
X
X
X X
X X
X
X
X X X X
X X X X
X
X X
X
X X
X
X
X X X
X X X
X X X
X X X
X
X
X X X X
X
X X X
X
X
0.004 0.087 0.008 0.016 0.023 0.009 0.058 0.027 0.018 0.016
0.012 0.207 0.008 0.038 0.007 0.256 0.156 0.115 0.053
0.031
0.000
0.011
0.214 0.125 0.069 0.039 0.031 0.026 0.025 0.025
0.164 0.093 0.051 0.024 0.015 0.010 0.007 0.006
0.287 0.255 0.210 0.135 0.101 0.041 0.019 0.004
X
X
X
0.914 0.938 0.952 0.788 0.890 0.928 0.953 0.957 0.529 0.676 0.801 0.896 0.920 0.949 0.960 0.968
0.054 0.036 0.018 0.142 0.070 0.040 0.023 0.012 0.257 0.199 0.129 0.065 0.048 0.025 0.016 0.007
0.032 0.027 0.031 0.070 0.040 0.032 0.025
0.081 0.078
ternary diagrams at a constant temperature and pressure for a series of organic acids, as shown in Figure 2, comparing the results for the phase equilibrium behavlor at 313 K and 15 MPa for acetic, propionic, and n-butyric acid. As can be seen in Figure 2, the change In hydrocarbon chain length for this series of straight-chain organic acids has the following effects on the phase equilibrium behavior. 1. The distribution coefficient of the organic acid between the supercritical fluid and aqueous phases changes gradually: it is less than 1 (the acid prefers the aqueous phase) for acetic acid, approximately equal to 1 (tie lines parallel to the COPwater side of the triangular diagram) for n-propionic acid, and greater than 1 (the acid preferentially distributes itself to the supercritical phase) for n-butyric acid. 2. The extent of the two-phase envelope (on a mole fraction basis) is less the higher the chain length of the acid. A striking similarity in the qualitative behavior of the three acid systems is observed when the ternary digrams shown In Figure 2 are compared with ternary diagrams for the systems acidwater-toluene at atmospheric pressure and 333 K from Alessi et al. (25). I t thus appears that, at high pressures, the solvent behavlor of carbon dioxide is similar to that of a nonpolar liquid
0.122 0.080
0.798 0.842
acetic acid
3
coz X X X X
X X
0.001
0.983 0.706 0.991 0.976 0.939 0.984 0.686 0.817 0.867 0.931 0.989 0.550 0.653
0.739 0.841 0.884 0.950 0.974 0.991
propionic acid
P = 15 MPG
A
P = 15 MPo
n-butyric acid
carbon diox:de
water
Flgure 2. Comparison of the ternary phase equilibrium behavior at 313 K and 15.0 MPa for acetic, propionic, and n-butyric acids.
326
Journal of Chemical and Engineering Data, Vol. 33, No. 3, 7988
Table IV. Experimental Results for the System Water-n-Butyric Acid-Carbon Dioxide at 313.1 K mole fractions phase 2
phase 1
P, MPa 2.06 2.04 2.06 1.96 1.99 2.03 2.03 3.96 3.96 3.97 3.98 4.05 3-phase 3.96 6.05 6.05 5.93 5.97 3-phase 5.89 6.95 7.10 6.99 3-phase 6.94 7.90 3-phase 7.93 7.88 9.90 9.74 10.12 9.97 15.10 15.01 15.10 19.73 19.40 19.60
H,O 0.039 0.106 0.297 0.467 0.685 0.877 0.932
0.942
0.952
coz
acid 0.756 0.683 0.554 0.446 0.268 0.105 0.056
H20
0.040
0.018
0.020
0.028
0.018
0.020
0.961 0.960 0.934 0.958 0.961 0.966 0.948 0.954 0.960 0.952 0.954 0.963
0.018 0.020 0.046 0.022 0.018 0.013 0.031 0.025 0.019 0.024 0.024 0.015
0.021 0.020 0.020 0.020 0.020 0.021 0.021 0.021 0.022 0.024 0.022 0.022
0.030 0.072 0.231 0.418 0.638 0.703 0.022 0.050 0.162 0.320 0.382 0.036 0.017 0.106 0.130 0.019 0.048 0.164
Table V. Pure Component Parameters and Physical Property Data Used
T,K kg m-3 313.2 992.1 333.2 983.3 acetic acid 313.2 1028.4 333.2 1006.0 propionic acid 313.2 967.8 n-butyric acid 313.2 938.0 carbon dioxide 313.2 333.2 component water
106b, kPa J m3 mol-2 m3 mol-' 7.38 0.8162 16.07 19.92 0.7912 15.98 4.69 2.431 51.13 12.09 2.376 51.45 1.34 3.534 67.95 0.39 4.778 84.36 0.3876 26.65 0.3696 26.65 PVP!
phase 3 CO2
0.205 0.211 0.148 0.089 0.047 0.018 0.012
0.963
P,
acid
a,
hydrocarbon solvent. The experimental data were correlated as described in the previous section. I n Table V, we present the pure component properties used for the correlation and the resulting equationof-state parameters. I n Table VI, the binary interaction parameters used in the model are given. These parameters were obtained by regressing binary phase equilibrium data. For the organic acid-carbon dioxide systems, no suitable binary data were available in the literature. For these systems, we used our own ternary data, extrapolated to zero water concentration, to regress the binary interaction parameters. The model was successful in predicting the qualitative characteristics of the experimentally observed behavior for all the cases studied, including the presence and approximate location of the three-phase equilibrium regions. The quantitative agreement between model predictions and experimental data is good in most cases, especially for the low-pressure region. For the high-pressure region, the model usually overpredicts the extent of the two-phase region and does not accurately reproduce the location of the plait points. However, since only binary parameters regressed from binary data that contain no
0.603 0.572 0.502 0.409 0.261 0.223 0.406 0.409 0.403 0.356 0.332 0.311 0.273 0.314 0.321 0.114 0.123 0.389
0.367 0.356 0.267 0.173 0.100 0.074 0.571 0.541 0.435 0.324 0.286 0.654 0.711 0.580 0.549 0.868 0.829 0.447
HzO
acid
c02
X
X
X
X
X
X
X
X
X
X X X X X X
X
X
X
X
X
X
X
X
X X
X X
X
X
X
X
X X
X
X X
X
X
X
X
X
X
X
X
X
X X
X X X
0.001
0.013
0.986
X X
X X
X
0.002 0.002
0.010 0.005
0.989 0.993
X
X
X
0.760 0.335 0.123 0.052 0.371 0.129 0.040 0.418 0.130 0.050
0.169 0.306 0.285 0.220 0.294 0.252 0.175 0.279 0.251 0.179
0.071 0.359 0.592 0.728 0.336 0.620 0.785 0.302 0.619 0.772
X
X
X X
X
Table VI. Interaction Parameters A12
binary system COz (1)-water (2) C02 (1)-acetic acid (2) COz (1)-propionic acid (2) C 0 2 (l)-n-butyric acid (2) water (1)-acetic acid (2) water (1)-propionic acid (2) water (l)-n-butyric acid (2)
=
421,
k , , = kzl J2 m3 313.2 0.024 -1740 333.2 0.028 -1710 530 all 0.005 0.011 381 313.1 0.023 527 313.2 -0.147 410 313.2 -0.142 333.2 580 1221 313.1 -0.127 -0.135 1686 313.2
T,K
information about the ternary system behavior were used, the overall performance of the model may be considered satisfactory. Reglstry No. CO,, 124-38-9; acetic acid, 64-19-7; propionic acid, 79-09-4; n-butyric acid, 107-92-6.
Llterature Clted (1) Paulaitls, M. E.; Krukonis, V. J.; Kurnik, R. T.; Reid, R. C. Rev. Chem. Eng. 1983, 1 , 181-250. (2) McHugh, M. A.; Krukonis. V. J. Supercrtical Fluid Extraction: Frinciples and Practice; Butterworths, Boston, 1986. (3) Paulaltls, M. E.; Gilbert, M. L.; Nash, C. A. Presented at the 2nd World Congress of Chemical Engineering, Montreal, Canada, 198 1. (4) McHugh, M. A,; Mallett. M. W.; Kohn, J. P. Presented at the 1981 annual AIChE meeting, New Orleans, LA, Nov. 9, 1981. (5) Gilbert, M. L.; Paulaltls, M. E. J . Chem. Eng. Data 1988, 37, 296-298. (6) Kuk, M. S.; Montagna, J. C. In Chemical Engineering af Supercritical Fluid Cond/f&ns; Paulaitls et al., Eds.; Ann Arbor Science: Ann Arbor, MI, 1983; Chapter 4, pp 101-111. (7) Radosz, M. Ber. Bunsen-Ges. f h y s . Chem. 1984, 88, 859-862. (8) Paulaltis, M. E.; Kander, R. G.; DlAndreth, J. R. Ber. Bunsen-Ges. Phys. Chem. 1984, 88. 869-875.
J. Chem. Eng. Data 1988, 33,327-333 (9) DiAndreth, J. R.; Paulaitis, M. E., Am. Chem. SOC.,Div. Fuel Chem. Pfepf. Pap. 1085, 30, 57-61. (IO) Fleck, R. N., Ph.D. Dissertation, Universlty of California, Berkeley, CA, 1967. (11) Panagiotopoulos, A. 2.: Reid, R. C. ACS Symp. Ser. 1987, 329, 1 15-1 29. (12) Panagiotopoulos, A. 2.: Reid, R. C. Fluid Phase Equllib. 1086, 29, 525-534. (13) Snedeker, R. A. Ph.D. Dissertation, Princeton University, Princeton, NJ, 1955. (14) Elgin, J. C.: Weinstock, J. J. J. Chem. Eng. Data 1950, 4, 3-12. (15) Francis, A. W. J. Phys. Chem. 1954, 58, 1099-1114. (16) Shimshick, E. J. CHEMTECH 1983, 13, 374-375. (17) Panagiotopoulos, A. 2. Ph.D. Dissertation, Massachusetts Institute of Technology, Cambridge, MA, 1986. (18) Wiebe, R.: Gaddy, V. L. J . Am. Chem. Soc. 1941, 63, 475-77. (19) Wiebe, R. Chem. Rev. 1041, 29,475-81.
327
(20) Katayama, T.; Oghaki, K.; Maekawa, G.; Goto, M.: Nagano, T. J. Chem. Eng. Jpn 1975, 8 , 89-92. (21) Peng, D-Y.; Robinson, D. 6. Ind. Eng. Chem. Fundam. 1976, 75, 59-64. (22) Panagiotopoulos, A. 2.; Kumar, S. Fluid Phase Equili6. 1985, 22, 77-88. (23) Reid, R. C.: Prausnitz, J. M.; Sherwood, T. K. The Properties of Gases and Liquids, 3d ed.; McGraw-Hili: New York, 1977. (24) WillSon, R. C. Ph.D. Dissertation, Massachusetts Instlute of Technology, Cambridge, MA, 1988. (25) Alessi, P.;Kikic, I.; Fermeglia, M. Fluid Phase Equll. 1984, 18, 93.
Received for review August 3, 1987. Accepted January 22, 1988. Financial support for this work was provided by the National Science Foundation, under Grant CPE-8318494.
Measurement and Model Prediction of Solubilities of Pure Fatty Acids, Pure Triglycerides, and Mixtures of Triglycerides in Supercritical Carbon Dioxide Thomas Bamberger,+ John C. Erickson,” and Charles
L. Cooney
Biochemical Process Engineering Center, Massachusetts Institute of Technology, Cambridge, Massachusetts 02 739
Sanat K. Kumars Department of Chemical Engineering, Massachusetts Institute of Technology, Cambridge, Massachusetts 02 739
The solubilities in CO, of three pure triglycerides (trilaurin, trimyristin, and tripalmitin) and their corresponding fatty acids (lauric acid, myristic acid, and palmitic acid) were measured at 313 K and at pressures between 8 and 30 MPa. The data were correlated by using a lattice model equation of state. Solubilities of members of a homologous series (fatty acid or triglyceride) decreased with Increasing molecular weight In accordance with previous findings. SoluMlltles of triglyceride mixtures in CO, were also measured. The solubility of the most soluble compound in the mixture was the same as the pure component solubility of that compound, but the solubilities of the less soluble species in mixtures were enhanced by the presence of more soluble triglycerides.
I ntroduction Edible oils are’ normally produced by extracting lipids from biological materials (such as soybeans) using n-hexane as a solvent. The high flammability and adverse health affects of residual solvent in food products make hexane a poor choice from an industrial standpoint. Because of these issues, supercritical fluids (SCF), especially supercritical carbon dioxide which is both nontoxic and nonflammable, are now being considered as replacements for hexane ( 7 , 2). SCF extraction of lipids has been studied extensively in recent years (2- 75)and a more comprehensive review has been prepared by McHugh and Krukonis ( 7 ) . The first reported commercial process using SCF for the fractionation of lipids was the Solexol process developed by the
’ Author to whom all correspondence should be addressed. Present address: lnstitut fur Thermische Verfahrenstechnik der Universitat Karlsruhe, Kaiserstr. 12, D-7500 Karlsruhe 1, West Germany. tPresent address: IBM Almaden Research Laboratory, Dept K91, 650 Harry Rd. San Jose, CA 95120.
M. W. Kellogg Co. in 1947 (76, 77). Supercriticalpropane was used to extract and fractionate edible oils in this process. Six plants were built to refine fish oil, animal fat, and vegetable oils
(17). More recently, Eisenbach (7)has described the separation of eicosapentenoic acid ethyl ester (C20:5)from a mixture of fatty acid ethyl esters derived from codfish oil using CO, at a constant temperature (323 K) and pressure (15 MPa). All of the fatty acid ethyl esters present in the mixture are soluble in CO, to some extent, but at the beginning of the extraction, the lower molecular weight esters are extracted preferentiaily. After most of the light esters have been stripped away, the desired product is then extracted in fairly high purity. Using a two-step process, Eisenbach achieved a product purity of approximately 92% with a yield of about 80%. While processes for the fractionation of oils have been tested, not much is known about the physical chemistry of lipid solubilii in supercriticalfluids. Most of the available information deals with bulk solubiliiies of complex oils in SCF, with liile data on the solubilities of the individual components of these mixtures. The only published quantitative data on the solubilities of pure fatty acids and triglycerides were measured by Chrastil (5)who reported the individual CO, solubilities of the following lipids: stearic acid, oleic acid, behenic acid, tributyrin, tripalmitin, tristearin, triolein, and trilinolein in the temperature range of 313-353 K and at pressures between 8 and 25 MPa. I n the present work, the solubilities of pure fatty acids and pure triglycerides were measured. The fatty acids used were lauric, myristic, and palmitic acids. The triglycerides were trilaurin (LLL), trimyristin (MMM), and tripalmitin (PPP). A recently developed lattice equation of state (EOS)(78) was used successfully to correlate the experimental data by using a single adjustable parameter per binary. The following triglyceride mixtures were also studied: LLL-MMM-CO,, LLL-PPP-C02, MMM-PPP-CO,, and LLL-MMM-PPP-CO,. All measurements were made at 313 K and at pressures between 8 and 30 MPa.
0021-9568/88/1733-0327$01.50/00 1988 American Chemical Society
